1
0
Fork 0
mirror of https://github.com/retspen/webvirtcloud synced 2026-11-10 20:26:48 +00:00
Commit graph

4 commits

Author SHA1 Message Date
catborise
0c9a002ef4 fix(networks): a network's addresses are read as libvirt writes them, and its changes keep each other
Reviewed with Codex against libvirt's network format and driver.

- Addresses: wvmNetwork found a network's <ip> by its text, so one
  written as <ip .../> or <ip><dhcp>... without whitespace had none,
  and a second address of a family overwrote the first. ip_networks
  reads every <ip> in order, its index being libvirt's parentIndex:
  family (ipv4 when written), address, subnet, DHCP (a <dhcp>, also
  with fixed addresses only), first range and fixed addresses. The
  parentIndex was guessed from the number of families, wrong when IPv6
  comes first; changes now name the address by its index.
- One read: the page read the network's XML once per getter, ten or
  more round trips; it reads it once (a cached_property of the
  request's wvmNetwork). A change starts from the saved definition
  read anew (VIR_NETWORK_XML_INACTIVE), not the running one: two QoS
  changes on a running network no longer undo the first. The XML
  editor shows and changes the saved definition, and defines this
  network only (its name, and its UUID if given), read for that with
  no entity expanded and handed to libvirt as written; an XML declaring
  its encoding is taken. The actions go back to the network's page by
  its own URL.
- Fixed addresses are found by the address they had (a hidden field),
  not by the editable box: changing it and pressing Delete deleted
  another entry or none. A change deletes the entry and adds it anew
  from a copy, keeping what the page does not show (a client id, a
  lease), and puts it back if the add fails: libvirt's own change
  finds an entry by any of its MAC, address or name, so it could
  replace another entry, and never by an IPv6 DUID. The address must
  be of the subnet, not the host's, and neither it, the MAC or DUID
  nor the name another entry's; an entry gone meanwhile is refused,
  not added anew; an unchanged one writes no audit row. They can be
  added while the network is stopped; a running network saved with
  other addresses than it runs with is not changed live, as
  parentIndex would name another <ip> of it.
- An address without a netmask or prefix takes libvirt's (IPv4 its
  class's, IPv6 /64), for its checks and the overlap check.
- The DHCP range must be in the subnet, in order and not take in the
  host's address.
- The create form takes a known mode only ("bogus" made an isolated
  network, a missing one a server error); NAT and routed networks need
  an IPv4 subnet, an isolated one may have none, a bridge or macvtap
  network needs its bridge or interface and gets no addresses; IPv6
  only with its box (whose value was "false", so never); no name or
  subnet of another network of the host. "Fixed Addresses" at create
  goes: it made a random-MAC entry for every address of the range,
  for no real VM.
- Refusals of a fixed address, the DHCP range, the XML and a QoS
  removal are messages, not server errors.
- The IPv6 forwarding of a bridge or macvtap network is the outside
  network's, not ROUTE.
- get_networks_info reads each network's XML once and gives its
  autostart and addresses too. Unused can_pxe, DHCP lease helpers,
  get_uuid and the per-family range and host getters go.
2026-10-29 10:09:14 +03:00
catborise
b24746c1cb fix(logs): blocking a blocked user writes no row; one upload submit path
Follow-ups from the review of the audit log layers:

- Blocking a user who is already blocked, or unblocking one who is
  not, changes nothing and now writes no row, like any other edit
  that changes nothing.
- write_chunk takes the request instead of the user and an optional
  request; the view always had one, so the branch only tests used
  is gone and every chunk 0 goes through the same audited submit.
- Tests: a busy clone creates no job and its row carries the page's
  message; a later chunk that fails writes no submit row; libvirt
  refusing a network or pool delete is a failed row.
2026-10-23 14:23:32 +03:00
catborise
d5df187de4 fix(networks, storages): a name in use is refused by the page again
The create forms of networks and storage pools checked the new name
against the host's list of networks or pools, but that list holds one
dict per object, so the name never matched and the check never fired;
libvirt refused the duplicate with its own wording. The check now
compares names: the page shows its own message, nothing is sent to
libvirt, and the Logs page gets a failed row "...: add: name already
in use".
2026-10-23 14:07:47 +03:00
catborise
5b8793ad4c feat(logs): changes to networks, storage, interfaces, secrets and filters are logged
The Logs page now records every change an admin makes to a compute's
libvirt objects, with the compute's name in the host column:

- networks: add, start, stop, delete, autostart, DHCP range, fixed
  addresses, QoS and XML edits;
- storage pools: add, start, stop, delete, autostart, from web and
  API; volumes: add and delete (web and API), clone and ISO upload;
- host interfaces: add, start, stop, delete;
- secrets: add, set value (never the value), delete;
- network filters: add, clone, delete, XML edits, rules and
  references.

A refused or failed change is a failed row with its reason (libvirt
text through safe_text). The volume clone and upload jobs write their
row when they end; a refused submission is a failed row at once. A
volume row names the volume libvirt created (a dir pool adds .qcow2 or
.img). Volume create or delete through the API on an inactive pool
(409) is a failed row.

Network filters: the rows read "Network filter f1: add" instead of
"f1 network filter is created", and their host is the compute's name
instead of its hostname. A duplicate name or UUID no longer raises a
bare exception, and a filter XML that cannot be parsed no longer ends
in UnboundLocalError.
2026-10-23 13:30:21 +03:00